Site considerations

If all System nodes are deployed at the same physical site, then events that affect any nodes at that site affect the whole system.

Single-site deployments are used often because all data is accessed at that site.

In this case, outages that affect the system affect the reader and writer of the data as well. It decreases the impact of correlated failures that affect only availability. Correlated failures that impact the survivability of data are no less critical.

Data loss is most likely due to a power outage that affects an entire rack or site. It can cause a loss of data that is written within the past minute, which is not yet saved to the drive. To avoid data loss from outages; conditioned, surge protected, and battery-backed up data is a must.

Alternatively, to gain the most in terms of resiliency and reliability out of a System, geographically disperse the Slicestor nodes across different data centers.